A new pathway to generative artificial intelligence by minimizing the maximum entropy
This addresses the limitations of current generative AI models for applications requiring data efficiency and controllable generation, representing a paradigm shift rather than an incremental improvement.
The paper tackles the problem of data-hungry and inflexible generative AI models by introducing a physics-driven framework that minimizes and maximizes entropy via adversary training to find informative yet low-noise data representations, resulting in a model that outperforms variational autoencoders and allows customization without retraining.
Generative artificial intelligence revolutionized society. Current models are trained by minimizing the distance between the produced data and the training set. Consequently, development is plateauing as they are intrinsically data-hungry and challenging to direct during the generative process. To overcome these limitations, we introduce a paradigm shift through a framework where we do not fit the training set but find the most informative yet least noisy representation of the data simultaneously minimizing the entropy to reduce noise and maximizing it to remain unbiased via adversary training. The result is a general physics-driven model, which is data-efficient and flexible, permitting to control and influence the generative process. Benchmarking shows that our approach outperforms variational autoencoders. We demonstrate the methods effectiveness in generating images, even with limited training data, and its unprecedented capability to customize the generation process a posteriori without any fine-tuning or retraining